^{Page 80 note 1} There were in 1928, 36,623 deaths from tuberculosis of all kinds, including 29,799 classified as respiratory and 6824 as due to all the other varieties of the disease. Statistical Review, Text, p. 9 (1928).

^{Page 80 note 2} There is of course another view which, without denying the influence of the amelioration of social conditions, attributes the decline of tuberculosis very largely to an increasing immunisation of the population which has resulted from the rise of industrialism and the growth of large cities, and which has been brought about by the consequent more frequent opportunities for infection pari passu with the increased density of the population.

Paradoxical though this view may seem at first sight it is rendered more reasonable by the fact, which was somewhat late in being recognised, that small doses of bacteria do not produce progressive disease but tend to immunise. The rule seems to be of general application, and it is probable that a community is resistant to its indigenous diseases, in proportion to their prevalence and the density of the population. It may be argued in favour of this thesis that it is in harmony with the general law that to all disturbing innovations nature tends to make corresponding adjustments.

^{Page 81 note 1} Standardised rates. See Statistical Review, Text, pp. 49, 50 (1921)Google Scholar and Pt I, Tables, p. 35 (1928).

^{Page 86 note 1} For the latter I have consulted the Preliminary fieport on the Causes of the Recent Decline in Tuberculosis Mortality written by Yves M. Biraud, M.D., for the League of Nations and published at Geneva in 1925. For the German figures I am indebted to Dr Stevenson of the General Register Office, Somerset House. They are taken from Die Bewegung der Bevölkerung (1922 and 1923) and Die Ursachen der Sterbefälle (1920 to 1923).

^{Page 87 note 1} New figures, based throughout on total population and therefore free from these defects have recently been published. But they are for both sexes combined and the male and female mortalities are not given separately. See Statistical Review, Pt I, Tables, p. 35 (1928)Google Scholar.

^{Page 89 note 1} For the figures on whioh these diagrams are based see Statistical Review, Text, pp. 49, 50 (1921)Google Scholar, and subsequent Reviews.

^{Page 92 note 1} See Statistical Review, Text, Table XXXIX, p. 53 (1927)Google Scholar.

^{Page 93 note 1} Registrar-General's Report, p. lxxv (1919)Google Scholar. See also Statistical Review, Text, p. 55 (1921)Google Scholar. Dr Stevenson's inferences Dr Stevenson's inferences were based on the mortality of females as being for that period more reliable than that of both sexes combined. But there is no reason to suppose that if we had the true figures for males these inferences would not be found to apply to them also.

^{Page 94 note 1} Lancet, i, p. 845 (1922)Google Scholar.

^{Page 94 note 2} Brit. Med. Journ. 2, 868 (1925)CrossRefGoogle Scholar.

^{Page 94 note 3} From Ergebnisse der Todesursachenstatistik im Deutschen Reiche, 1914–19Google Scholar.

^{Page 99 note 1} They numbered about 18,000 (see Table V, p. 96).

^{Page 100 note 1} Since this was written I have heard a graphic account given, quite spontaneously, by a sailor of the state of hunger produced by the war in neutral Norway, and of how, though the people were enriched, the children in the ports there would scramble and fight like wild beasts for a biscuit if it were thrown among them.

^{Page 101 note 1} Sixth Ann. Rep. Board of Control for 1919, p. 29Google Scholar. Tuberculosis of all kinds is chosen rather than respiratory tubercle because the figures are more accessible, but it makes little difference since 90 per cent, of the tubercle deaths in the asylums were from phthisis.

^{Page 102 note 1} Fifth Ann. Rep. Board of Control for 1918, p. 24Google Scholar. In their Report for 1917 the Board had said that there was “a fairly general consensus of opinion amongst the Medical Superintendents… that the food restrictions…were accompanied by a lowering of the weight of patients and a general loss of nutrition.” This they thought “must have induced a lessening of vital resistance, which would go far to explain the increased liability to infection in the case of such communicable diseases as dysentery and tuberculosis,” loc. cit. p. 23.

^{Page 102 note 2} In the Sixth Ann. Rep. Board of Control for 1919, p. 35Google Scholar: the importance of a dietary which includes a sufficiency of such food accessories, as animal fats—notably milk and butter— “from lack of which the vulnerability to tuberculosis is admittedly enhanced” was clearly recognised.

^{Page 102 note 3} Green, and Mellanby, , Brit. Med. Journ. 2, 691 (1928)CrossRefGoogle Scholar; and Report on Accessory Food Factors, published by the Med. Res. Committee, 1919.